Circuit Court of Appeals, Second Circuit.
*267 Morris Hirsch, Oscar W. Jeffery, and Dean, Fairbank, Hirsch & Foster, all of New York City, for appellant.
Watson, Bristol, Johnson & Leavenworth, of New York City (Charles P. Bauer, of New York City, Leon Robbin, of Washington, D. C., and Laurence Bristol, of New York City, of counsel), for appellees.
Before L. HAND, SWAN, and CHASE, Circuit Judges.
L. HAND, Circuit Judge.
The plaintiffs sued the defendant in Brooklyn upon patent No. 1,891,207; the defendant counterclaimed upon two patents, Nos. 1,815,768 and 1,789,949. The judge passed a decree holding the plaintiffs' patent valid and infringed and dismissing the counterclaim. The defendant appealed. We need discuss only the plaintiffs' patent, because we dispose of the counterclaim on other grounds. All the patents are for improvements in dry electrolytic condensers; the defendant's are the same as those concerned in a suit brought by it against the Concourse Electric Company and decided by this court in 65 F.(2d) 386; we may refer to what we said there by way of general background for our discussion. The plaintiffs' patent is for another electrolyte, the viscous compound with which the fabric is saturated which separates the anode and the cathode plates, and which serves to restore the film upon the anode if it is broken through. It does not too much simplify the facts to say that the invention consisted only in substituting glycol in the place of glycerin in the electrolyte; for the other ingredients, boric acid and ammonium borate, were concededly old in the art. Glycol has undoubtedly generally superseded glycerin since the plaintiffs introduced it and was an advance in the art; moreover, it was new in a dry electrolyte when Ruben filed his application on June 19, 1930. But it had been disclosed as part of the electrolyte of a wet condenser twice before; first, by Engle in patent No. 1,672,714, application for which was filed on August 29, 1927; and again by Fansteel, in a British patent No. 319,033 of 1929, the application for which was filed on May 15, 1928. Both disclosures mention glycol as an optional substitute for glycerin, obviously regarding the two as interchangeable and readily suggested alternatives. And indeed although the claims in suit prescribe glycol, other claims are for glycerin in its place, and the specifications make it plain that though glycol was preferred, glycerin would serve. Thus at page 1, lines 78-87: "In the invention the viscous material is not confined to ethylene glycol for glycerine and various other glycols or glycerols are applicable; after many tests, however, ethylene glycol was found to give the best results. Electrolyte compositions prepared with the use of glycerols in lieu of glycols, with the proportions of the ingredients remaining substantially the same, appear to be thicker, that is, to have a lower degree of flowability." It was apparently because of its greater "flowability" that Ruben preferred glycol. The substance had been known for long in laboratories, but its large-scale production only dates from about 1921, within a few years after which it began to arouse interest in its practical uses, as appears from several papers which were put in evidence describing its possibilities. It may be assumed to have become cheap enough for use in such arts as these not very long before Ruben filed his application, though the date does not definitely appear in the record. (The bills in evidence do not bear out the conclusion that there was a very substantial fall in its value after July, 1930.) We know just how Ruben came upon it. He was a very prolific inventor and he directed an assistant, Raines, to carry on a series of parallel experiments in electrolytes with glycerin and with glycol; Raines found that the glycol had a better "leakage characteristic," and the art has confirmed his conclusions.
In the light of all this it seems to us that the patent does not rest upon an authentic invention, but upon one of those steps in an art which demand *268 only patient experiment. Especially in chemistry it is possible to proceed by a system of trial and error, varying formulas by permutation and combination, and recording the results of each. Much that is valuable has been so discovered, and we will not say that the profitable survivals from such elimination can never be inventions; salvarsan for example, as its other name, "606," indicates, was hit upon by this method. Ordinarily invention demands more than that; some resumption of a line of experiment from which the art had looked away [E. I. Dupont De Nemours & Co. v. Glidden Co., 67 F.(2d) 392, 397 (C. C. A. 2)], some departure which required originality or independence of conception; something more than routine testing of obvious combinations. Here we can find nothing more. The defendant itself tried out glycol and abandoned the experiment, because uncertain of the merchantable quality available and because its customers were reasonably satisfied with glycerin. Indeed, Georgiev's own electrolyte patent No. 1,815,768 suggested glycol as a possible substitute six months later, and though he also included glucose which was not satisfactory, the suggestion shows that the mere conception was commonplace enough. Such invention as there was, lay in the verification. While it is always the safest course to test a putative invention by what went before and what came after, it is easy to be misled. Nothing is less reliable than uncritically to accept its welcome by the art, even though it displace what went before. If the machine or composition appears shortly after some obstacle to its creation, technical or economic, has been removed, we should scrutinize its success jealously; if at about the same time others begin the same experiments in the same or nearby fields, or if these come to fruition soon after the patentee's, the same is true. Such a race does not indicate invention. We should ask how old was the need; for how long could known materials and processes have filled it; how long others had unsuccessfully tried for an answer. If these conditions are fulfilled, success is a reliable touchstone; but success in the circumstances at bar proves nothing. The patent is invalid.
There remain the Aerovox patents on which we passed in the former case. Our disposition of these requires a statement of the earlier relations between the parties, and in a little detail. The plaintiffs sued the defendant on June 3, 1931, upon two other Ruben patents, each for electrolytic condensers. The parties came to an accommodation before trial and signed some papers on May 2, 1932. The defendant took a license under the two patents, and got an option for six months after its appearance upon any patent which might issue upon the then pending application for patent No. 1,891,207 that which we have just discussed. It never took up that option. Thus the defendant was safe from prosecution under all three of the plaintiffs' patents if it chose; but it must pay royalties. However, that was not all. The defendant then held the two Georgiev patents which are now the subject of the counterclaim, and some arrangement as to these also was desired; it took the form of an exchange of letters between it and the Mallory Company, by which the defendant promised not to sue "for infringement of any United States patent now owned or controlled by us, including specifically the Georgiev United States patents Nos. 1,789,948 and 1,815,768 * * * by reason of the manufacture, use or sale of condensers under the Ruben patents Nos. 1,710,073 and 1,714,191, in accordance with the practices followed by you or them at any time from January 1, 1929, to date in the manufacture of such condensers as you or they have sold in substantial quantities." The Mallory Company's letter was in the same terms, and protected the defendant against the Ruben patents then in suit. True, it does not appear that the defendant was at the time claiming infringement against the Mallory Company, but that makes no difference; the purpose was not to disturb that company in exploiting the Ruben patents already issued, so long as it confined itself to what it had been doing. The Mallory Company was in fact making condensers with the same electrolyte that the defendant now claims to be one of the infringements of the Georgiev electrolyte patent No. 1,815,768, and, as we understand it, it acknowledges the plaintiffs' immunity so far, though it somewhat faintly suggests that the same electrolyte when used in a new container is not protected. There is no merit in this; the letter gave immunity against any infringement of Georgiev's electrolyte patent and it was as much infringed by an electrolyte in a cardboard cover, as in an aluminum. However, the plaintiffs thereafter made two changes in the electrolyte; first, they decreased the percentage of boric acid, and, second, they reduced the boiling point. Their process *269 on May 2, 1932 the "A" solution had been to add thirty ounces of boric acid to 200 c.c. of ammonium hydroxide and 640 c.c. of ethylene glycol. Less than a minute after this had been slowly brought to a boil, eighteen more ounces of boric acid was added and the boiling continued until the boiling point was 125° C. In the "B" solution the addition of the second boric acid was omitted; the "C" solution was like the "B" except that the boiling point was 120°, instead of 125°. From no figures at hand can we say whether a reduction of three-eighths of the boric acid brought the solution nearer to the Georgiev proportions than before; we have not the specific gravity of the substances. But that is not necessary. The letter should be read with another agreement of even date giving to the Mallory Company a license whenever in its "opinion" it should adopt "a construction electrolyte or method which * * * comes within the scope of any of the claims" of the Georgiev patents. This made the company judge for that purpose of whether it infringed. That of course was not final on the issue of infringement; it might choose to stand suit rather than accept a license; but this latitude presupposed that the immunity given by the accompanying letter was not to be construed rigidly. Both documents were parts of a single compromise, and it was no doubt expected that the Mallory Company would take a license whenever it thought that it infringed by any change. The immunity would end where the license began, if the company lived up to its engagements fairly. It is unlikely that in such a settlement the purpose was to hold it to the very letter of its then practices, and the reduction of three-eighths in one element ought not to forfeit the protection so granted. So much for the "B" solution. The "C" was further from Georgiev's electrolyte because the boiling point was lower. His contribution, as the defendant insistently proclaims, lay not in the mix but in the boiling; and this indeed helps to confirm our conclusion as to the "B" solution also, in which the boiling point remained the same as in the "A." The infringements of this patent, if there were any, were and are excused.
The case as to the cover patent depends upon whether the Mallory Company had begun to sell aluminum condensers in "substantial quantities" on May 2, 1932. For the most part its condensers had been sheathed in cardboard which concededly did not, and could not, infringe; Georgiev's specifications only mentioned metal "cans" enclosing the condenser proper. On the other hand, it must be owned that the Mallory Company had not sold the aluminum can condenser in large quantities on May 2, 1932; its records show that it had sold only 2,007, and nearly 200,000 cardboard condensers had been sold during about the same period. The defendant makes a not very ingenuous effort further to reduce this number by showing that most of them may have been samples, but it is entirely apparent that the witness did not mean to say that the proportion of these was greater than in the case of other condensers. We recur to the obvious purpose underlying the whole compromise; the Mallory Company was to be allowed to exploit its own two patents without interference; the Aerovox Company its two patents. There were indeed limits to this; no departures from, and no innovations upon, what had gone before were to be allowed, but that was all. Just as this was not intended as a strait-jacket upon variations in factory practice, so it seems to us it did not demand extensive marketing. The "line" must have gone into real production; should have ceased to be merely experimental, but that was enough. This is especially true as applied to this patent, because unless 2007 was a "substantial quantity," the letter gave the Mallory Company no protection whatever against it, though it was expressly mentioned as one of the two which should cause it no trouble. This covers condenser No. 2 in the record; Nos. 4, 9, and 10 indeed appeared later, but as they were substantially the same except in size, they fall within the reasoning as to solutions "B" and "C."
We do not forget that in an earlier form the letters had read differently; the condensers to be protected were the "type" or "types" then being manufactured, a suggestion of the Mallory Company, not acceptable to the defendant. But we cannot agree that the rejected form necessarily covered more than that adopted. That might be true if the parties had changed from "type or types of condenser now being manufactured," to "condensers now being manufactured". They did not; they substituted a phrase which had little relation to either form, and which was as general as the first. The permitted "practices" included not only those then in use, but any which had been abandoned within three years. It was a *270 plan of substantial peace barring the pending application leaving each side free to carry on its business on existing lines. Such a plan cannot be realized in a niggardly spirit, by catching at minor variations of factory practice. Had such a power been meant to be held in reserve, the purpose ought to have been more clearly expressed. The decree dismissing the counterclaim is affirmed.
Decree reversed as to patent to Ruben, No. 1,891,207, and bill dismissed for invalidity; decree affirmed as to patents to Georgiev, Nos. 1,789,949 and 1,815,768, because the defendant had given the plaintiffs immunity as to all the alleged infringements. The costs will be divided.
